[
https://issues.apache.org/jira/browse/FLINK-32266?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17729668#comment-17729668
]
Martijn Visser commented on FLINK-32266:
----------------------------------------
[~xiechenling] It's a fair argument that this is too complex, but it's not a
bug. Improving this is planned as part of
https://cwiki.apache.org/confluence/display/FLINK/FLIP-246%3A+Multi+Cluster+Kafka+Source
> Kafka Source Continues Consuming Previous Topic After Loading Savepoint
> -----------------------------------------------------------------------
>
> Key: FLINK-32266
> URL: https://issues.apache.org/jira/browse/FLINK-32266
> Project: Flink
> Issue Type: Bug
> Components: Connectors / Kafka
> Affects Versions: 1.15.3
> Environment: Flink version: 1.15.3
> Kafka Connector version: 1.15.3 FLIP-27
> Reporter: xiechenling
> Priority: Major
>
> I encountered an issue with the Flink Kafka Connector's Kafka Source where it
> continues consuming data from a previously consumed topic even after loading
> a savepoint and configuring it to consume data from a different topic.
>
> Steps to reproduce:
> # Set up the Kafka Source to consume data from Topic A.
> # Start the Flink job.
> # Stop the job and create a savepoint.
> # Modify the configuration to consume data from Topic B.
> # Load the job from the savepoint and start it.
> # Observe that the job consumes data from both Topic A and Topic B, instead
> of just Topic B.
>
> Expected behavior:
> After loading a savepoint and configuring the Kafka Source to consume data
> from a new topic, the job should only consume data from the newly configured
> topic.
>
> Actual behavior:
> The Kafka Source continues consuming data from the previous topic (Topic A),
> in addition to the newly configured topic (Topic B).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
